Contactless Weight Monitoring of Grow-out Nile Tilapia in a Recirculated Aquaculture System Using Multiple Linear Regression Supervised Machine Learning Approach

Abstract

Fish weight is the most important characteristic in aquaculture, conservation, fisheries, and management because it is related to the growth of each fish in a specific area.

Through the studies presented, this study aims to develop a contactless weight monitoring of grow-out Nile-tilapia in a recirculated aquaculture system using the parameters being supervised such as Ammonia, Temperature, and Total Dissolved Solids (TDS).  This can help the fish farmer to analyze and determine the fish growth and fish weight without actual contact with the fish. With the integration of the Internet of Things (IoT) in this research, it will be a more secure and reliable system that is in demand nowadays. A real-time and efficient monitoring system is really wanted for some critical parameters which can improve the value and determine future projection by using the previously saved data.

Starting from the water quality parameter detection index by the following sensors: pH, TDS, Temperature, and dissolved oxygen. Ammonia is supervised in a mathematical method by means of the temperature and pH level. This will be interpreted by the main program stored in Arduino UNO. The reading will be sent online and processed using the derived AI model then water quality parameter reading, and prediction results will be displayed in the designed interface on a laptop or PC which will provide information for fish farmers as their guide to take necessary action.

Results showed that the use of IoT in monitoring water quality parameters in a recirculating aquaculture system was a real-time and efficient technique.  As evidence of the efficiency of the developed system, there is a 98% precision of the prediction model for the fish weight. Therefore, the developed a system that can be a viable method to provide contactless fish weight monitoring which will be helpful in the fishery industry research.

Index Terms—Fish weight, IoT, Linear regression, RAS, TDS.